IBM eserver pSeries sizing and capacity planning : a practical guide / / [G. Benton Gibbs ... et al.]

• Record Nr.: UNINA9910828466903321
• Pubbl/distr/stampa: [S.l.], : IBM, International Technical Support Organization, c2004
• Edizione: [1st ed.]
• Descrizione fisica: xviii, 546 p. : ill
• Collana: IBM redbooks
• Altri autori (Persone): GibbsG. Benton
• Disciplina: 004.2/4
• Soggetti: Computer capacity - Planning; IBM computers
• Lingua di pubblicazione: Inglese
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Note generali: The "e" in "eserver" is printed as the symbol for electronic.; "March 2004."
• Nota di bibliografia: Includes bibliographical references and index.
• Nota di contenuto: Front cover -- Contents -- Notices -- Trademarks -- Preface -- The team that wrote this redbook -- Become a published author -- Comments welcome -- Part 1 Introduction to sizing and capacity planning -- Chapter 1. Overview, concepts, and approach -- 1.1 Definitions of common terms -- 1.2 Concepts -- 1.2.1 Required knowledge and experience -- 1.2.2 Sizing with capacity planning -- 1.2.3 The sizing problem -- 1.2.4 Sizing inputs -- 1.2.5 Sizing outputs -- 1.2.6 Who performs sizing -- 1.3 Sizing and resizing process -- 1.3.1 System design and requirements -- 1.3.2 Sizing model -- 1.3.3 Hardware requirements -- 1.3.4 Building block choices -- 1.3.5 eConfig for the price -- 1.3.6 Sales, purchase, install, and production -- 1.3.7 Gathering performance data -- 1.3.8 Performance tuning -- 1.3.9 Estimated or measured growth -- 1.3.10 Capacity planning -- 1.3.11 Resizing model -- 1.3.12 rPerf reliance -- 1.4 Weighing sizing components -- 1.4.1 Memory size -- 1.4.2 Disk type and number -- 1.4.3 Adapters for disk, tape and network -- 1.4.4 Software -- 1.4.5 Summary -- 1.5 The importance of the right amount of information -- 1.5.1 Brain overload -- 1.5.2 Summary -- 1.6 A practical sizing method -- 1.6.1 Segmentation -- 1.7 Performance theory -- 1.8 General rules of thumb for RDBMS memory -- 1.8.1 Application resident set -- 1.8.2 RDBMS data and file system cache -- 1.8.3 RDBMS utilization rules of thumb -- 1.8.4 Utilization -- 1.8.5 RDBMS raw data ; to disk rules of thumb -- 1.8.6 RDBMS disk use rules of thumb -- 1.9 The performance saturation curve -- 1.10 Successive approximation and sizing levels -- 1.11 Plagiarism -- 1.12 Triangulation -- 1.12.1 A triangulation story -- 1.13 Common sizing mistakes -- 1.13.1 Sizing report outline -- 1.13.2 A sizing story -- 1.14 The eConfig configurator -- 1.14.1 Configurator test -- 1.15 Cost-based sizing method.; 1.16 High availability and disaster recovery -- 1.17 Capacity Upgrade on Demand -- 1.18 Sizing for Linux on pSeries -- Part 2 Components involved in sizing and capacity planning -- Chapter 2. Hardware components -- 2.1 Performance methodology -- 2.2 Overview of pSeries systems -- 2.2.1 Autonomic computing -- 2.2.2 e-business on demand -- 2.2.3 Reliability, availability, and serviceability features -- 2.2.4 Capacity Upgrade on Demand -- 2.3 pSeries processors -- 2.3.1 Processor descriptions -- 2.3.2 RISC/CISC concepts -- 2.3.3 Superscalar architecture: Pipelines and parallelisms -- 2.3.4 32-bit versus 64-bit computing -- 2.3.5 Performance of processors -- 2.3.6 Processor evolution -- 2.4 Memory -- 2.4.1 Memory hierarchy -- 2.4.2 Locality concept -- 2.4.3 Caches -- 2.4.4 Memory cycles -- 2.4.5 Virtual memory concepts -- 2.4.6 Memory affinity -- 2.4.7 Large page support -- 2.5 Input/output -- 2.5.1 Peripheral Component Interconnect -- 2.5.2 PCI-X -- 2.6 Storage architectures -- 2.6.1 Direct access storage -- 2.6.2 Storage area networks -- 2.6.3 Network-attached storage -- 2.6.4 RAID -- 2.6.5 IBM TotalStorage Enterprise Storage Server -- 2.6.6 IBM TotalStorage Fibre Array Storage Technology -- 2.6.7 IBM 7133 Serial Disk System -- 2.6.8 IBM TotalStorage Expandable Storage Plus 320 -- 2.6.9 The IBM TotalStorage Network Attached Storage -- 2.7 Additional hardware considerations -- 2.7.1 Multiprocessor configurations -- 2.7.2 NUMA -- 2.7.3 Logical partitioning -- 2.7.4 Dynamic logical partitioning (5.2.0) -- 2.7.5 Dynamic CPU sparing and CPU Guard (5.2.0) -- 2.7.6 UE-Gard (5.2.0) -- Chapter 3. Software components -- 3.1 AIX -- 3.1.1 History of AIX -- 3.1.2 AIX kernel -- 3.1.3 Modes of operation (execution modes) -- 3.1.4 AIX 5L kernel subsystems -- 3.1.5 Multitasking and multithreading support -- 3.1.6 64-bit kernel -- 3.2 Workload Manager -- 3.2.1 Classes.; 3.2.2 Tiers -- 3.2.3 Class attributes -- 3.3 Linux -- 3.3.1 Linux for pSeries -- 3.3.2 Linux and AIX -- 3.3.3 Logical partitioning -- 3.3.4 Other related information and links -- Chapter 4. Benchmarks -- 4.1 Introduction to benchmarks -- 4.2 OLTP benchmarks -- 4.2.1 TPC-C benchmark -- 4.3 Business intelligence benchmarks -- 4.3.1 TPC-H benchmark -- 4.4 e-business benchmarks -- 4.4.1 TPC-W benchmark -- 4.4.2 SPEC JBB2000 benchmark -- 4.4.3 SPECweb99 benchmark -- 4.5 High Performance Computing benchmarks -- 4.5.1 SPEC CPU2000 benchmark -- 4.5.2 LINPACK benchmark -- 4.6 ISV benchmarks -- 4.6.1 SAP Standard Application benchmarks -- 4.6.2 Oracle Applications Standard benchmark -- 4.6.3 Siebel platform sizing and performance program benchmark -- 4.7 Relative performance -- Part 3 Sizing pSeries systems -- Chapter 5. General sizing -- 5.1 Where to locate the Balanced System Guideline -- 5.2 Six golden sizing principles -- 5.2.1 Correct processor configuration -- 5.2.2 Balanced systems -- 5.2.3 CPU magic number calculations -- 5.2.4 Estimating CPU power -- 5.2.5 Estimating memory sizing -- 5.2.6 Estimating disk sizing -- 5.3 The Balanced System Guideline overview -- 5.3.1 Problems with sizing -- 5.3.2 Assumptions: Prerequisites for using the spreadsheet -- 5.3.3 Spreadsheets: Pros and cons -- 5.3.4 The Balanced System Guideline sections -- 5.4 The Balanced System Guideline details -- 5.4.1 Introduction sheet -- 5.4.2 Performance and ; balanced systems sheets -- 5.4.3 Balanced system examples -- 5.4.4 LPAR sheet -- 5.4.5 pSeries costs -- 5.4.6 Price-based sizing -- 5.4.7 Sizing new systems -- 5.4.8 Sizing CPU and RAM sheet -- 5.4.9 Sizing and planning disks sheet -- 5.4.10 Sizing Results sheet -- 5.4.11 Calibration sheet -- 5.4.12 Calibrating a new workload example: SAP, DB2, pSeries 650 -- 5.5 Resizing existing systems for upgrades -- 5.5.1 Assumptions.; 5.5.2 ResizeCPU sheet -- 5.5.3 ResizeRAM sheet -- 5.5.4 ResizeDisk sheet -- 5.5.5 ResizeDiskUse sheet -- 5.5.6 Modeling to add new workloads -- 5.6 Balanced System Guideline and sizing levels -- 5.6.1 Sizing for Level 2: 'Ball park' or rough estimates -- 5.6.2 RDBMS server sizer for level 3: Consider opinion -- 5.6.3 Sizing for Level 4: Sizing from measured data -- 5.7 Business intelligence sizing -- 5.7.1 Business intelligence golden rules -- 5.7.2 Business intelligence sizing approaches -- 5.7.3 Business intelligence sample configurations -- 5.8 Disk and stripe sizing -- 5.8.1 Disk sizing -- 5.8.2 Stripe sizing -- 5.9 pSeries 670 and 690 RIO-2 I/O Sizing Tool -- 5.9.1 Notes and assumptions -- 5.9.2 Readme sheet -- 5.9.3 Adapters sheet -- 5.9.4 Results sheet -- 5.9.5 p670/p690 errors sheet -- 5.9.6 RIO-2 loops sheet -- 5.10 Review and summary -- Chapter 6. Application-specific sizing -- 6.1 IBM applications -- 6.1.1 DB2 -- 6.1.2 Lotus Domino -- 6.1.3 Tivoli Storage Manager -- 6.1.4 WebSphere -- 6.2 ISV applications -- 6.2.1 [e-mail] sizing support -- 6.2.2 Quick e-sizing guides -- 6.3 IBM Eserver Sizing Guide -- 6.4 Network File System sizing -- 6.4.1 Functionality -- 6.4.2 Cache management on an NFS client -- 6.4.3 Performance considerations -- 6.4.4 Method and sizing factors -- Part 4 Capacity planning -- Chapter 7. AIX tools for data gathering -- 7.1 AIX standard tools -- 7.1.1 The vmstat command -- 7.1.2 The iostat command -- 7.1.3 The sar command -- 7.1.4 The svmon command -- 7.1.5 The ps command -- 7.1.6 The ipcs command -- 7.1.7 The topas command -- 7.2 Performance Toolbox -- 7.3 AIX Workload Manager -- 7.3.1 Configuring AIX Workload Manager -- 7.3.2 System capacity and sizing for workload management -- 7.3.3 The wlmstat command -- 7.4 Performance Management Services for AIX -- 7.4.1 Architecture -- 7.4.2 Utilization.; 7.4.3 Comparison, correlation, forecast -- 7.4.4 PM/AIX usage -- 7.4.5 Data collection -- 7.4.6 Thresholds -- 7.4.7 SRM reports -- 7.4.8 Executive reports -- 7.4.9 Capacity reports -- 7.4.10 Workload specific reports -- 7.4.11 Application response metric reports -- 7.4.12 System analysis and forecast with PM/AIX -- Chapter 8. Features and tools for capacity planning -- 8.1 Performance Toolbox -- 8.1.1 Tool utilization strategy -- 8.1.2 azizo -- 8.1.3 xmtrend -- 8.1.4 jazizo -- 8.1.5 wlmperf -- 8.2 Workload Manager -- 8.2.1 Typical UNIX system capacity sizing -- 8.2.2 Server consolidation considerations -- 8.2.3 System capacity sizing for workload management -- 8.2.4 Conclusion -- 8.3 Dynamic LPAR and CUoD -- 8.3.1 Configuration alternative -- 8.3.2 DLPAR benefit -- 8.3.3 Partitioning misconceptions -- 8.3.4 Example situations using LPAR -- 8.3.5 DLPAR sizing considerations -- 8.3.6 DLPAR and applications -- 8.3.7 CUoD advantage: Pay as you grow -- 8.3.8 Workload Manager versus DLPAR -- 8.3.9 Capacity planning for DLPAR -- 8.3.10 DLPAR examples -- 8.4 IBM Insight tools -- 8.4.1 IBM Insight for SAP R/3 overview -- 8.4.2 IBM Insight for Oracle database -- Part 5 Appendixes -- Appendix A. Sanity check before upgrading -- Identifying the workloads -- Setting objectives -- Identifying critical resources -- Minimizing critical-resource requirements -- Using the appropriate resource -- Reducing the requirement for the critical resource -- Structuring for parallel use of resources -- Reflecting priorities in resource allocation -- Repeating ; the tuning steps -- Applying additional resources -- Appendix B. Sample for CPU resource usage calculation -- Abbreviations and acronyms -- Related publications -- IBM Redbooks -- Other resources -- Online resources -- How to get IBM Redbooks -- Help from IBM -- Index -- Back cover.